breast cancer control: signs of progress, but more work required

ARTICLE IN PRESS

The Breast (2005) 14, 429–438

THE

BREAST

0960-9776/$ - sdoi:10.1016/j.b

�Tel.: +33 4 7E-mail addr

www.elsevier.com/locate/breast

ORIGINAL ARTICLE

Breast cancer control: Signs of progress, but morework required

Peter Boyle�

International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon Cedex08, France

KEYWORDSBreast cancer;Mortality

Summary Breast cancer represents 10% of the global cancer burden and there isno population, or population sub-group, which has yet been identified to be at a trulylow risk of developing the disease. Mortality rates have been steadily growing for

ee front matter & 2005reast.2005.10.001

273 8577.ess: [email protected].

nearly a century in many countries and it is only during the past decade that therehave been signs of a sustainable decrease in mortality rates in a number of western-lifestyle countries. This represents considerable progress in breast cancer controland, although different factors contribute to different degrees in differentcountries, is mainly due to increased breast awareness, earlier detection and thedelivery of the most appropriate therapy to women with the disease. The failure toprevent the incidence from continuing to rise represents to a great extent the failureto understand the precise mechanisms of breast carcinogenesis and the role of riskdeterminants whose alteration in society could lead to a reduced risk of developingthe disease. The declines seen in mortality represent a considerable success, butthere is no room for complacency until research can impact positively on reducingincidence.& 2005 Published by Elsevier Ltd.

Introduction

Breast cancer is a major Public Health issue world-wide. It is estimated1 that in 2002 there were1,151,298 new cases of breast cancer diagnosed,representing over 10% of the world’s cancerburden. This figure has risen, approximatelydoubled, from an estimated figure of 550,000 casesin 1975.2 In 2000, there were an estimated 410,712deaths caused by breast cancer and over 4.4 million

Published by Elsevier Ltd.

women living with breast cancer world-wide. In themore developed countries, there were 636,128incident cases compared with 514,072 in lessdeveloped countries. This translated into 189,765breast cancer deaths in the more developedcountries and 220,648 breast cancer deaths in lessdeveloped countries. This clearly indicates thatpatients may be at a more advanced stage atdiagnosis in less developed countries and that theremay be poorer treatment outcomes.

In Europe in 2004. it was estimated that therewere 360,746 incident cases of breast cancer,representing 13.01% of the total cancer burden,

www.elsevier.com/locate/breast

ARTICLE IN PRESS

Figure 1 Fifteen highest and 15 lowest age-standardisedincidence rates from breast cancer reported in the mid-1990s.

Scotland, 1911-2000

0

5

10

15

20

25

30

1911

-191

519

16-1

920

1921

-192

519

26-1

930

1931

-193

519

36-1

940

1941

-194

519

46-1

950

1951

-195

519

56-1

960

1961

-196

519

66-1

970

1971

-197

519

76-1

980

1981

-198

519

86-1

990

1991

-199

519

96-2

000

Breast Cancer Mortality in

Figure 2 Average-annual age-adjusted mortality rate per100,000 person—years from breast cancer in Scotland,1911–2000.

Mortality from Breast Cancer in United States, 1930-2000

0

5

10

15

20

25

30

35

40

1930

1934

1938

1942

1946

1950

1954

1958

1962

1966

1970

1974

1978

1982

1986

1990

1994

1998

Figure 3 Average-annual age-adjusted mortality rate per100,000 person—years from breast cancer in the UnitedStates, 1931–2000.

P. Boyle430

and 129,010 deaths from breast cancer.3 Thecumulative lifetime risk of developing breastcancer before the age of 75 was 7.8% and of dyingfrom the disease 2.1%.

The 15 highest incidence rates world-wide arecommonly found in North American populationswhile the lowest rates come from a variety ofpopulation groups in the low and medium resourcecountries (Fig. 1).4 Clearly, however, there is nopopulation that now records a truly low rate ofbreast cancer and there is no group of women whohave been identified as truly being at a low risk ofbreast cancer.

Trends in breast cancer mortality

Mortality data by age and time period have existedfor breast cancer in Scotland since 1911.5 The (age-adjusted) death rate from breast cancer has risenalmost continually, with the exception of the

wartime years, until falling from 1990 onwards(Fig. 2). Such a pattern has also been seen in theUnited States, with rates remaining fairly stablebetween 1930 and 1990 and subsequently experi-encing a decline (Fig. 3).

After increasing from 1951 to around 1990,mortality rates have been falling in most EuropeanUnion countries.6 Trends in some representativecountries (United Kingdom, Netherlands, Italy,France, Sweden) are presented in Fig. 4a and thegreater decline in the United Kingdom is quitenoticeable. In central and eastern European coun-tries the mortality rate is rising, as exemplifiedin Hungary, Poland and the Russian Federation(Fig. 4b). Although lower, mortality rates are risingin Hong Kong and Japan (Fig. 4b). In countries ofNorth (e.g. Canada and United States) and SouthAmerica (e.g. Uruguay and Argentina) breast cancermortality rates are similar to those in westernEurope and, once again, declining since around 1990after having increased slightly (Fig. 4c).

ARTICLE IN PRESS

0

5

10

15

20

25

30

35

United KingdomNetherlandsSwedenItalyFrance

0

5

10

15

20

25

30

35HungaryRussiaPolandHong KongJapan

0

5

10

15

20

25

30

35

AustraliaUnited StatesCanadaArgentinaUruguay

Breast Cancer Mortality, 1951-2000All ages, age-standardised rates per 100,000



(a)

(b)

(c)

1951

1954

1957

1960

1963

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

2003

1951

1954

1957

1960

1963

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

2003

1951

1954

1957

1960

1963

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

2003

Figure 4 All-ages, age-standardised mortality from breast cancer in women, 1951–2000. Selected countries in broadgeographical regions: (a) United Kingdom, France, Netherlands, Italy and Sweden, (b) Hungary, Russia, Poland, HongKong and Japan; and (c) Australia, United States, Canada, Argentina and Uruguay.

Breast cancer control: Signs of progress, but more work required 431

Why is breast cancer mortality declining?

Mammographic screening

Potential contributions to these observed reduc-tions in breast cancer mortality could come fromprimary prevention, chemoprevention, screening(secondary prevention), improved therapy (tertiaryprevention) or combinations of these. Contribu-tions to the decline in western European, Australianand American countries have come from a varietyof sources including widespread mammographicscreening, more precise diagnosis and more womenreceiving the optimal treatment for their condi-

tion, including notably the widespread use oftamoxifen. It is also unlikely that the same factorsexert the same impact in all countries withdeclining rates.

Despite a recent unfounded controversy,7 there issubstantial evidence that mammographic screeningfor breast cancer is effective in reducing themortality from the disease.8 For women aged50–64, mammographic screening in organised pro-grammes with quality control standards in placecan be recommended.9

In Scotland, the National Breast Cancer ScreeningProgramme for women aged 50–64 was introducedin the late 1980s and became effective in the

ARTICLE IN PRESS

0

20

40

60

80

100

12019

8119

8219

8319

8419

8519

8619

8719

8819

8919

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

01

102.9(100%)

76.5(74.3%)

aged 50-64 in Scotland, 1981-2000.Mortality from Breast Cancer in Women

Figure 5 Annual, age-adjusted, truncated (50–64) breastcancer mortality rate in Scotland, 1981–2001.

P. Boyle432

following years. By 1998, 71.1% of women in theage range were in compliance with the screeningguidelines, i.e., had been screened within theprevious three years. The Health of the NationTarget was to reduce the death rates for breastcancer in the population invited for screening byat least 25% by the year 2000 (Baseline 1990).Mortality data from the General Registrar Office forScotland show that the mortality rate has droppedfrom 102.9 per 100,000 in 1990 to 76.3 in 2000: areduction of 25.7% and within the initial target(Fig. 5). Screening can be transferred successfullyfrom the clinical trial arena to impact on breastcancer mortality at a population level.

The aim of mammographic screening is to reducethe mortality from breast cancer in the populationby early detection and effective treatment ofbreast cancer when it may be in a curable phase.The aim of mammographic screening is not toreduce the overall mortality rate among women,nor to reduce the incidence of the disease.

The impact of reduction in the stage at diagnosisof breast cancer has been outlined recently usingthe data on breast cancer patients from theSurveillance, Epidemiology and End Results (SEER)Program in a competing-risk analysis calculatingprobabilities of death from breast cancer and othercauses according to stage, race and age atdiagnosis.2 The probability of death from breastcancer after nearly 28 years of follow-up rangedfrom 0.03 to 0.10 for patients with in situ disease to0.70–0.85 for patients with distant disease, de-pending on race and age. Among patients withlocalised or regional disease and known ER status,the probability of death from breast cancer afternearly 11 years of follow-up ranged from 0.04 to0.11 for patients with localised ER-positive tumoursof 2 cm or less to 0.37–0.53 for patients withregional ER-negative tumours.

Chemoprevention

Findings about reductions in contralateral breastcancer in breast cancer patients in adjuvant trialsof tamoxifen led to the undertaking of four large-scale randomised trials of breast cancer preven-tion.11–15 Results on breast cancer reduction, majorside effects and mortality from the four tamoxifenprevention trials have been collected and over-viewed.16 For comparison results from the ralox-ifene trial in postmenopausal women withosteoporosis (MORE Trial)17 and 11 adjuvant breastcancer treatment trials18 using approximately fiveyears of tamoxifen have been considered simulta-neously.

Overall, the four tamoxifen prevention trialsindicate a 38% (28–46%) reduction in incidence ofbreast cancer. The reduction was confined to ER-positive tumours where it was 48% (36–58%) and aslight, although not significant, increase in ER-negative tumours was noted. Endometrial cancerwas increased to 2.4 (1.5–4.0) times the rate incontrols and the relative risk of venous thromboem-bolic events was 1.9 (1.4–2.7). Overall, there wasno effect on non-breast cancer mortality and theonly cause showing a mortality increase waspulmonary embolism.

The majority of cases of breast cancer are causedby hormonal factors, defining these in their broad-est sense. It has been known for over 100 yearsnow, since the work of Beatson in Glasgow19 thatbreast cancer was a disease related to hormones.However, it has been only recently that a hormonalintervention, using tamoxifen, has been shown toreduce the risk of breast cancer (and only ER-positive breast cancer).11–18 While this is greatnews in terms of proof-of-principle, the level ofside effects produced by tamoxifen, some serious interms of the medical threat and others in terms ofthe impact on quality of life, rule this out for use inthe general population of women at present,despite the protection offered.

Primary prevention

While mortality rates from breast cancer have beenfalling in Scotland, the incidence has been con-tinuing to increase (Fig. 6a). This increase is notcompletely due to any over-diagnosis of in situlesions produced by screening (Fig. 6b) since thebiggest contribution to the increased incidencecomes from an increase in invasive breast cancer.

While the reductions taking place in breastcancer mortality rates are very gratifying andrewarding, the lack of a reduction in the incidence

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

80

90

1951

1954

1957

1960

1963

1966

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

Incidence

0

10

2030

40

50

6070

80

90

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

Incidence of Invasive and in situBreast Cancer in Scotland, 1975-2000

Incidence and Mortality from Breast

Cancer in Scotland, 1951-2000

Mortality

In situ

Invasive

(a)

(b)

Figure 6 All-ages, age-adjusted (a) incidence (1975–

2000) and mortality (1951–2000) rates per 100,000 fromBreast Cancer; and (b) incidence of invasive and in situbreast cancer in Scotland (1975–2000).


rates of breast cancer in most countries isextremely disappointing. Despite much researchactivity, little progress is being made in identifyingrisk determinants for breast cancer, which could bemodifiable and lead to a reduction in breast cancerrisk.

Aetiology of breast cancer

Identifying effective strategies and interventions toprevent breast cancer remains one of the greatchallenges in oncology today. Although women whohave first-degree relatives with a history of breastcancer are at increased risk, a major pooledanalysis reveals that most will never develop breastcancer, and most who do will be aged over 50 whentheir cancer is diagnosed. In countries where breastcancer is common, the lifetime excess incidence ofbreast cancer is 5.5% for women with one affectedfirst-degree relative and 13.3% for women withtwo. Eight out of nine women who develop breastcancer do not have an affected mother, sister, ordaughter.20

Pregnancy- and hormone-related factors

While many risk factors have been associated withaltered breast cancer risk, there are remarkablyfew risk determinants, the sub-group of risk factorswhose alteration would directly alter the risk ofbreast cancer.21 The majority of cases of breastcancer are caused by hormonal factors, definingthese in their broadest sense. Oestrogens, whetherendogenous or exogenous, whether direct exposureto oestrogens or to chemicals with oestrogenicproperties, hold the key to determining breastcancer risk in women. The subject of breast cancerrisk factors such as hormonal factors (age atmenarche, age at first birth, parity), exogenoushormonal sources (oral contraceptives and hor-mone replacement therapy) or dietary intakes (fatintake) has been well studied and some risk factorshave been clearly identified, some of which aretheoretically subject to alteration. Breast cancerrisk factors have been comprehensively reviewedpreviously22 although there have been some ad-vances in the understanding of breast cancer causesand their contribution to potential strategies forprevention.

Hormone levels normally change throughout awoman’s life for a variety of reasons. Many suchhormonal changes occur during pregnancy, changesthat may influence a woman’s chances of developingbreast cancer later in life. As a result, over severaldecades a considerable amount of research has beenand continues to be conducted to determine the roleof pregnancy-related factors as influences on awoman’s chance of developing breast cancer laterin life. Associations with breast cancer risk, age atfirst birth and parity were pioneered by the work ofMacMahon.23 Currently, there is well-establishedevidence that early age at first term birth is relatedto lifetime decrease in breast cancer risk; increasingparity is associated with a long-term risk reduction,even when controlling for age at first birth; theadditional long-term protective effect of young ageat subsequent term pregnancies is not as strong asfor the first term pregnancy; a nulliparous womanhas approximately the same risk as a woman with afirst term birth around age 30; breast cancer risk istransiently increased after a term pregnancy;induced abortion is not associated with an increasein breast cancer risk; recognised spontaneous abor-tion is not associated with an increase in breastcancer risk; and, long duration of lactation providesa small additional reduction in breast cancer riskafter consideration of age at and number of termpregnancies.

The lack of or short lifetime duration ofbreastfeeding typical of women in developed

ARTICLE IN PRESS

P. Boyle434

countries makes a major contribution to the highincidence of breast cancer in these countries. Acollaborative analysis24 estimated that that the riskof breast cancer was significantly decreased by4.3% for every 12 months of breastfeeding inaddition to a 7% decrease for each birth. Breastfeeding practices are potentially modifiable andcould usefully be promoted as a preventativestrategy.

The hypothesis that an interrupted pregnancymight increase a woman’s risk of breast cancer wasexamined in a prospective, population-based co-hort study in Denmark.25 After adjustment forknown risk factors, induced abortion was notassociated with an increased risk of breast cancer(relative risk, 1.00; 95% confidence interval [CI],0.94–1.06). Following this, the Collaborative Groupon Hormonal Factors in Breast Cancer has broughttogether the worldwide epidemiological evidenceon the possible relation between breast cancer andprevious spontaneous and induced abortions.26

Pregnancies that end as a spontaneous or inducedabortion were found not to increase a woman’s riskof developing breast cancer. Collectively, thestudies of breast cancer with retrospective record-ing of induced abortion yielded misleading results,possibly because women who had developed breastcancer were, on average, more likely than otherwomen to disclose previous induced abortions.26

Exogenous hormones

There appear to be two main conclusions regardingthe association between breast cancer risk and useof oral contraceptives. First, while women aretaking combined oral contraceptives and in the 10years after stopping there is a small increase in therelative risk of having breast cancer diagnosed.Second, there is no significant excess risk of havingbreast cancer diagnosed 10 or more years afterstopping use. The cancers diagnosed in women whohad used combined oral contraceptives were lessadvanced clinically than those diagnosed in womenwho had never used these contraceptives: for ever-users compared with never-users.27

The risk of having breast cancer diagnosed isincreased in women using HRT and increases withincreasing duration of use.10 This effect is reducedafter cessation of use of HRT and has largely, if notwholly, disappeared after about 5 years. Thesefindings should be considered in the context of thebenefits and other risks associated with the use ofHRT.28

Current use of hormone-replacement therapy(HRT) increases the incidence of breast cancer.

The Million Women Study was set up to investigatethe effects of specific types of HRT on incident andfatal breast cancer. In current users of each type ofHRT the risk of breast cancer increased withincreasing total duration of use. Ten years’ use ofHRT is estimated to result in 5 (95% CI 3–7)additional breast cancers per 1000 users of oestro-gen-only preparations and 19 (15–23) additionalcancers per 1000 users of oestrogen–progestogencombinations. Use of HRT by women aged 50–64years in the UK over the past decade has resulted inan estimated 20,000 extra breast cancers, 15,000associated with oestrogen–progestogen; the extradeaths cannot yet be reliably estimated. It could beconcluded that current use of HRT is associatedwith an increased risk of incident and fatal breastcancer; the effect is substantially greater foroestrogen-progestogen combinations than for othertypes of HRT.29

Anthropometric indices and physical activity

There is a clear association between anthropo-metric indices and the risk of breast cancer. Usingpooled data from seven prospective cohort studies,comprising 337,819 women and 4385 incidentinvasive breast cancer cases in total, after control-ling for reproductive, dietary and other riskfactors, the pooled relative risk (RR) of breastcancer per height increment of 5 cm was 1.02 (95%CI: 0.96, 1.10) in pre-menopausal women and 1.07(95% CI: 1.03, 1.12) in postmenopausal women.Body mass index (BMI) showed significant inverseand positive associations with breast cancer amongpre- and postmenopausal women, respectively.30

These findings indicate that height is an indepen-dent risk factor for postmenopausal breast cancerwhile in pre-menopausal women, this relation isless clear.30

Using data from 73,542 pre-menopausal and103,344 postmenopausal women from nine Eur-opean countries taking part in the EPIC study,relative risks (RRs) were calculated, using mea-sured height, weight, BMI and waist and hipcircumferences adjusted for study centre, ageand other risk factors.31 In postmenopausal women,current HRT modified the body size-breast cancerassociation. Among non-users, weight, BMI and hipcircumference were positively associated withbreast cancer risk (p trend p0.002); obese women(BMI430) had a 31% excess risk compared withwomen with BMIo25. Among HRT users, bodymeasures were inversely but non-significantly asso-ciated with breast cancer. Excess breast cancer riskwith HRT was particularly evident among lean

ARTICLE IN PRESS


women. Pooled RRs per height increment of 5 cmwere 1.05 (95% CI 1.00–1.16) in pre-menopausaland 1.10 (95% CI 1.05–1.16) in postmenopausalwomen. In postmenopausal women not takingexogenous hormones, general obesity is a signifi-cant predictor of breast cancer, while abdominalfat assessed as waist–hip ratio or waist circumfer-ence was not related to excess risk when adjustedfor BMI. Among pre-menopausal women, weightand BMI showed non-significant inverse associationswith breast cancer.31

Increased physical activity appears to be inver-sely related to risk of breast cancer,32 althoughthere remains some inconsistency in the findingsprobably attributable to limitations in methodsused to assess physical activity. A recent analysis ofthe Nurses’ Health Study II revealed no overallassociation between physical activity and risk ofbreast cancer among pre-menopausal women, butsuggests that the effect of physical activity couldbe substantially modified by the underlying degreeof adiposity.33 Physical activity, and weight control,are potentially modifiable risk factors that can berecommended at present34 although there could bea more substantial benefit arising from furtherresearch.

Dietary factors

The association between breast cancer risk anddietary factors has been a source of great interestfor many years and also, it must be said, a source ofconsiderable frustration for all concerned. Apartfrom the measurement problems, results haveoften been contradictory and the situation is hardlyclear at the present time.

International comparisons and case-control stu-dies have suggested a positive relation betweendietary fat intake and breast cancer risk, butprospective studies, most of them involving post-menopausal women, have not supported thisassociation. The question of whether intakes ofspecific types of fat are associated with breastcancer risk independently of other types of fat wasinvestigated in a pooled analysis of eight prospec-tive studies.35 The pooled RRs (95% CI) for anincrement of 5% of energy were 1.09 (1.00–1.19)for saturated, 0.93 (0.84–1.03) for monounsatu-rated and 1.05 (0.96–1.16) for polyunsaturated fatcompared with equivalent energy intake fromcarbohydrates. For a 5% energy increment, theRRs were 1.18 (95% CI 0.99–1.42) for substitutingsaturated for monounsaturated fat, 0.98 (95% CI0.85–1.12) for substituting saturated for polyunsa-turated fat and 0.87 (95% CI 0.73–1.02) for

substituting monounsaturated for polyunsaturatedfat. No associations were observed for animal orvegetable fat intakes. These associations were notmodified by menopausal status and are consistentwith only a weak positive association with substitu-tion of saturated fat for carbohydrate consump-tion.35 Using the same source, no significantassociations were found between intake of meator dairy products and risk of breast cancer.36

However, most recently, in the Nurses’ HealthStudy II, intake of animal fat, mainly from redmeat and high-fat dairy foods, during pre-meno-pausal years was associated with an increased riskof breast cancer.37

No material association among 88,678 women(aged 34–59 years at baseline) in the Nurses’ HealthStudy was observed between carbohydrate intake,glycaemic index and glycaemic load, total dietaryfibre intake, and breast cancer risk.38 When thishad been examined among 90,655 pre-menopausalwomen in the Nurses’ Health Study II (aged 26–46years in 1991) associations between carbohydrateintake or glycaemic load and breast cancer riskamong young adult women appeared to differ bybody weight. The data do not support a strongassociation between fibre intake and breast cancerrisk.39

Results from the pooled analysis of eight pro-spective studies suggest that fruit and vegetableconsumption during adulthood is not significantlyassociated with reduced breast cancer risk.40 Whendata from five prospective studies were combinedto compare the death rates from common diseasesof vegetarians with those of non-vegetarians withsimilar lifestyles, in comparison with regular meateaters, mortality from ischaemic heart disease was20% lower in occasional meat eaters, 34% lower inpeople who ate fish but not meat, 34% lower inlacto-ovovegetarians, and 26% lower in vegans.There were no significant differences betweenvegetarians and non-vegetarians in mortality fromcerebrovascular disease, stomach cancer, color-ectal cancer, lung cancer, breast cancer, prostatecancer, or all other causes combined.41

Alcohol consumption

To assess the risk of invasive breast cancerassociated with total and beverage-specific alcoholconsumption and to evaluate whether dietary andnon-dietary factors modify the association, datafrom six prospective studies were examined.Alcohol consumption was associated with a linearincrease in breast cancer incidence in women overthe range of consumption reported by most women.

ARTICLE IN PRESS

P. Boyle436

Among women who consume alcohol regularly,reducing alcohol consumption is a potential meansof reducing breast cancer risk.42

In a collaborative re-analysis of a large number ofstudies, and analyses based on 58,515 women withinvasive breast cancer and 95,067 controls from 53studies, the findings suggest that about 4% of thebreast cancers in developed countries are attribu-table to alcohol. In developing countries, wherealcohol consumption among controls averaged only0.4 g per day, alcohol would have a negligible effecton the incidence of breast cancer. Smoking haslittle or no independent effect on the risk ofdeveloping breast cancer; the effect of alcohol onbreast cancer needs to be interpreted in thecontext of its beneficial effects, in moderation,on cardiovascular disease and its harmful effects oncirrhosis and cancers of the mouth, larynx, oeso-phagus and liver.43

Cigarette smoking, frequently considered inaetiological studies together with alcohol consump-tion, does not appear to be related to breast cancerrisk.44

Environmental exposures

An increased risk of breast cancer among womenwho have been exposed to ionising radiation,particularly around puberty, has been widelyaccepted45 even when the exposure has been tolow doses.46 Attention has recently increasinglyfocused on environmental exposures and environ-mental chemical pollutants as risk factors forbreast cancer in women. The common nature ofbreast cancer in women, and the common nature ofsuch exposures make such an association plausiblealthough far from proven as causal. Also, given theresistant nature of concentrations in adiposetissue, it is important to attempt to identify anyrisk resulting from this exposure.

Based on current evidence, the risk of breastcancer associated with exposure to organochlorinepesticides and their residues appears to be small, ifit exists at all. However, one caveat is that theeffect of individual or groups of polychlorinatedbiphenyls (PCBs) as risk factors for breast cancermerits some further investigation. Indeed, a clearmessage is that great care and attention must betaken to any statistical analysis of such chemicalexposures given the potential for co-linearity thatexists between the various congeners.47 Theseresults indicate that studies of PCB congeners andhealth require an in-depth statistical analysis inorder to better understand the complex issuesrelated to their co-linearity.47

Environmental exposure to organochlorines hasbeen examined as a potential risk factor for breastcancer. In 1993, five large US studies of womenlocated mainly in the north-eastern United Stateswere funded to evaluate the association of levelsof 1,1-dichloro-2,2-bis(p-chlorophenyl) ethylene(DDE) and PCBs in blood plasma or serum withbreast cancer risk.48 We present a combinedanalysis of these results to increase precision andto maximise statistical power to detect effectmodification by other breast cancer risk factors.Combined evidence does not support an associationof breast cancer risk with plasma/serum concen-trations of PCBs or DDE. Exposure to thesecompounds, as measured in adult women, isunlikely to explain the high rates of breast cancerexperienced in the north-eastern United States.48

Chronic exposure to electro-magnetic fields(EMF) has been hypothesised to increase femalebreast cancer risk. Although no studies with directmeasures have been reported, household electricalappliances frequently provide the highest magneticfield exposure levels in any residence, with electricblankets among the strongest sources. However,the most recent data available suggest thatconcerns that exposure to EMFs from domesticappliances may increase breast cancer risk appearto be unfounded. This does not rule out thepresence of any risk and further studies, somehowwith more direct measures of EMF exposure wouldbe welcomed. However, it can be concluded thatwithin our current knowledge, domestic exposuresto EMF do not present a major source of breastcancer risk.47

Discussion

The great advance in recent years has been thedecline seen in the mortality rate from breastcancer in many (developed) countries even thoughthe incidence appears still to be increasing. This isa success for breast awareness, treatment ad-vances and early detection programmes and thedecline would be intensified if interventions couldbe identified and implemented that would lead torisk reduction and a fall in incidence. However,successful primary prevention still appears to besomewhat far away within our current knowledgeof the causes of the disease.

The majority of cases of breast cancer are causedby hormonal factors, defining these in their broad-est sense. It has been known for over 100 yearsnow, since the work of Beatson in Glasgow19 thatbreast cancer was a disease related to hormones.

ARTICLE IN PRESS


However, it has been only recently that hormonalintervention, using tamoxifen, has been shown toreduce the risk of breast cancer (only ER-positivebreast cancer).16

In contrast to well-established pregnancy-relatedrisk factors, there appears to be a clear gulfbetween the surety with which these risk factorsare understood and knowledge of potential me-chanisms to explain these associations. There are alarge number of epidemiologic gaps including anabsence of knowledge of the mechanism wherebypregnancy at an early age protects against breastcancer; whether pregnancy and age at pregnancymodify radiation-induced breast cancer risk; theeffects of age at pregnancy on sub-groups ofwomen (e.g. those with BRCA-1 and BRCA-2mutations); the mechanism by which lactationaffects breast cancer risk; the temporal patternof breast cancer risk following lactation; the effectof lactation on women with BRCA-1 and BRCA-2mutations; the effect of gender of offspring on risk;the impact of birth weight of offspring on breastcancer risk; the impact of multiple births in thesame pregnancy, with and without assisted repro-ductive technology; the breast cancer risk implica-tions of abnormal pregnancies (e.g. spina bifida,late foetal death, fertility treatment-inducedpregnancy, pre-term delivery, small for gestationalage offspring); the mechanism by which pre-eclampsia reduces breast cancer risk; the distinc-tion between hypertension and pre-eclampsia withrespect to breast cancer risk; and, whether gesta-tional diabetes is associated with breast cancerrisk.

As far as information from animal experimentsgoes, few items are well established: pregnancyprotects against subsequent chemical carcinogen-induced breast cancer in rats and mice; oestrogenand progesterone combinations protect againstcarcinogen-induced cancer in rodents by mimickingpregnancy; and, short-term oestrogen exposure, atlevels of oestrogen mimicking pregnancy, is pro-tective for carcinogen-induced cancer in rats.

The gaps in our understanding of breast carcino-genesis remain enormous. The mechanisms ofhormone action when they are given before orafter chemical carcinogen exposure are not under-stood. The relationship between pregnancy and riskof pre-neoplastic lesions is also not understood andknowledge is absent regarding the levels, determi-nants and interactions of pregnancy-related mam-motrophic factors, ligands, and receptors. It isclearly important to increase knowledge of themechanisms of hormonal carcinogenesis. It iscertain that endogenous hormonal factors play acritical role in determining a woman’s risk of

developing breast cancer; yet, the precise hor-mone(s) involved, their critical levels and theircritical timing remain unknown. It is important toobtain a clearer understanding of the hormonalmechanisms of breast carcinogenesis. It is impos-sible to alter the age at menarche, to advisewomen to have an earlier pregnancy, to advisethem to have more births and to avoid having birthsat a later age. Instead, a clearer understanding ofthe precise hormonal mechanisms would open uppossibilities for effective chemoprevention thatcould lead to significant reductions in the incidenceof and death from this common disease.

References

1. Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2002: cancerincidence, mortality and prevalence worldwide. IARCCancerBase No. 5, version 2.0. Lyon: IARC Press; 2004.

2. Parkin DM, Sternsward J, Muir CS. Estimates for 1957.3. Boyle P, Ferlay J. Cancer incidence and mortality in Europe,

2004. Ann Oncol 2005;16:481–8.4. Case RAM, Coghill C, Davies JM, Harley JL, Hytten CA,

Pearson JT, et al. Serial mortality tables. Neoplasticdiseases, vol 4. Scotland, 1911–1970. Deaths and deathsrates by sex, age, site and calendar period. London:Institute of Cancer Research; 1976.

5. Parkin DM, Whelan SL, Ferlay J, Storm H. Cancer incidencein five continents. IARC CancerBase No. 7. Lyon: IARC Press;2005.

6. Boyle P. Mammographic screening: after the dust hassettled. Breast 2003;12:351–6.

7. Boyle P, d’Onofrio A, Maisonneuve P, Severi G, Robertson C,Tubiana M, et al. Measuring progress against cancer InEurope: has the 15% decline targeted for 2000 come about?Ann Oncol 2003;14:1312–25.

8. IARC. Handbook of cancer prevention, no. 7. Breast cancerscreening. Lyon: IARC Press; 2002.

9. Boyle P, Autier P, Bartelink H, Baselga J, Boffetta P, Burn J,et al. European code against cancer and scientific justifica-tion: third version (2003). Ann Oncol 2003;14:973–1005.

10. Schairer C, Mink PJ, Carroll L, Devesa SS. Probabilities ofdeath from breast cancer and other causes among femalebreast cancer patients. J Natl Cancer Inst 2004;96:1311–21.

11. Cuzick J, Forbes J, Edwards R, et al. First results from theInternational Breast Cancer Intervention Study (IBIS-I): arandomised prevention trial. Lancet 2002;360:817–24.

12. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen forprevention of breast cancer: report of the National SurgicalAdjuvant Breast and Bowel Project P-1 Study. J Natl CancerInst 1998;90:1371–87.

13. Powles TJ, Eeles R, Ashley S, et al. Interim analysis of theincidence of breast cancer in the Royal Marsden Hospitaltamoxifen randomised chemoprevention trial. Lancet 1998;352:98–101.

14. Veronesi U, Maisonneuve P, Costa A, et al. Prevention ofbreast cancer with tamoxifen: preliminary findings from theItalian randomised trial among hysterectomised women.Lancet 1998;352:93–7.

15. Veronesi U, Maisonneuve P, Sacchini V, et al. Tamoxifen forbreast cancer among hysterectomised women. Lancet2002;359:1122–4.

ARTICLE IN PRESS

P. Boyle438

16. Cuzick J, Powles T, Veronesi U, Forbes J, Edwards R, AshleyS, et al. Overview of main outcomes in breast cancerprevention trials. Lancet 2003;361:296–300.

17. Cauley JA, Norton L, Lippman ME, et al. Continued breastcancer risk reduction in postmenopausal women treatedwith raloxifene: 4-year results from the MORE trial. BreastCancer Res Treat 2001;65:125–34.

18. EBCTCG-Early Breast Cancer Trialists’ Collaborative Group.Tamoxifen for early breast cancer: an overview of therandomised trials. Lancet 1998;351:1451–67.

19. Beatson GT. On the treatment of inoperable cases ofcarcinoma of the mammae: suggestions of new methods oftreatment, with illustrative cases. Lancet 1896;ii:162–5.

20. Collaborative Group on Hormonal Factors in Breast Cancer.Familial breast cancer: collaborative reanalysis of individualdata from 52 epidemiological studies including 58,209women with breast cancer and 101,986 women withoutthe disease. Lancet 2001;358:1389–99.

21. MacMahon B. A biologic framework for the risk factors forbreast cancer. Adv Oncol 1994;10:2–10.

22. Henderson BE, Pike MC, Bernstein L, Ross RK. Breast cancer.In: Schottenfeld D, Fraumeni JF, editors. Cancer epidemiol-ogy and prevention. 2nd ed. New York: Oxford UniversityPress; 1996.

23. MacMahon B. General motors cancer research prizewinnerslaureates lectures. Charles S. Mott Prize. Reproduction andcancer of the breast. Cancer 1993;71:3185–8.

24. Collaborative Group on Hormonal Factors in Breast Cancer.Breast cancer and breastfeeding: collaborative reanalysis ofindividual data from 47 epidemiological studies in 30countries, including 50302 women with breast cancer and96973 women without the disease. Lancet 2002;360:187–95.

25. Melbye M, Wohlfahrt J, Olsen JH, Frisch M, Westergaard T,Helweg-Larsen K, et al. Induced abortion and the risk ofbreast cancer. N Engl J Med 1997;336:81–5.

26. Collaborative Group on Hormonal Factors in Breast Cancer.Breast cancer and abortion: collaborative reanalysis of datafrom 53 epidemiological studies, including 83,000 womenwith breast cancer from 16 countries. Lancet 2004;363:1007–16.

27. Collaborative Group on Hormonal Factors in Breast Cancer.Breast cancer and hormonal contraceptives: collaborativereanalysis of individual data on 53.297 women with breastcancer and 100.239 women without breast cancer from 54epidemiological studies. Lancet 1996;347:1713–27.

28. Collaborative Group on Hormonal Factors in Breast Cancer.Breast cancer and hormone replacement therapy: colla-borative reanalysis of data from 51 epidemiological studiesof 52,705 women with breast cancer and 108,411 womenwithout breast cancer. Lancet 1997;350:1047–59.

29. Beral V. Million women study collaborators. Breast cancerand hormone-replacement therapy in the Million WomenStudy. Lancet 2003;362:419–27.

30. Van den Brandt PA, Spiegelman D, Yaun SS, Adami HO,Beeson L, Folsom AR, et al. Pooled analysis of prospectivecohort studies on height, weight, and breast cancer risk. AmJ Epidemiol 2000;152:514–27.

31. Lahmann PH, Hoffmann K, Allen N, van Gils CH, Khaw KT,Tehard B, et al. Body size and breast cancer risk: findingsfrom the European Prospective Investigation into Cancer AndNutrition (EPIC). Int J Cancer 2004;111:762–71.

32. IARC. Handbook on cancer prevention, Weight control andphysical activity, vol 6. Lyon: IARC Press; 2002. p. 144–54.

33. Colditz GA, Feskanich D, Chen WY, Hunter DJ, Willett WC.Physical activity and risk of breast cancer in premenopausalwomen. Br J Cancer 2003;89:847–51.

34. Boyle P, Autier P, Bartelink H, et al. European code againstcancer and scientific justification: third version (2003). AnnOncol 2003;14:973–1005.

35. Smith-Warner SA, Spiegelman D, Adami HO, Beeson WL, vanden Brandt PA, Folsom AR, et al. Types of dietary fat andbreast cancer: a pooled analysis of cohort studies. Int JCancer 2001;92:767–74.

36. Missmer SA, Smith-Warner SA, Spiegelman D, Yaun SS, AdamiHO, Beeson WL, et al. Meat and dairy food consumption andbreast cancer: a pooled analysis of cohort studies. Int JEpidemiol 2002;31:78–85.

37. Cho E, Spiegelman D, Hunter DJ, Chen WY, Stampfer MJ,Colditz GA, et al. Premenopausal fat intake and risk ofbreast cancer. J Natl Cancer Inst 2003;95:1079–85.

38. Holmes MD, Liu S, Hankinson SE, Colditz GA, Hunter DJ,Willett WC. Dietary carbohydrates, fiber, and breast cancerrisk. Am J Epidemiol 2004;159:732–9.

39. Cho E, Spiegelman D, Hunter DJ, Chen WY, Colditz GA,Willett WC. Premenopausal dietary carbohydrate, glycemicindex, glycemic load, and fiber in relation to risk ofbreast cancer. Cancer Epidemiol Biomarkers Prev 2003;12:1153–8.

40. Smith-Warner SA, Spiegelman D, Yaun SS, Adami HO, BeesonWL, van den Brandt PA, et al. Intake of fruits and vegetablesand risk of breast cancer: a pooled analysis of cohortstudies. JAMA 2001;285:769–76.

41. Key TJ, Fraser GE, Thorogood M, Appleby PN, Beral V, ReevesG, et al. Mortality in vegetarians and nonvegetarians:detailed findings from a collaborative analysis of 5 prospec-tive studies. Am J Clin Nutr 1999;70(3 Suppl):516S–24S.

42. Smith-Warner SA, Spiegelman D, Yaun SS, van den Brandt PA,Folsom AR, Goldbohm RA, et al. Alcohol and breast cancer inwomen: a pooled analysis of cohort studies. JAMA1998;279:535–40.

43. Collaborative Group on Hormonal Factors in Breast Cancer.Alcohol, tobacco and breast cancer—collaborative reana-lysis of individual data from 53 epidemiological studies,including 58 515 women with breast cancer and 95 067women without the disease. Br J Cancer 2002;87:1234–46.

44. International Agency for Research on Cancer (IARC) Mono-graphs on the Evaluation of Carcinogenic Risks to humans.Tobacco smoke and involuntary smoking, vol. 83. Lyon: IARCPress; 2004.

45. Boice JD, Stone BJ. Interaction between radiation and otherbreast cancer risk factors. Late biological effects of ionizingradiation, vol. 1. Vienna: IAEA; 1978. p. 231–49.

46. Modan B, Chetrit A, Alfandary E, Katz L. Increased risk ofbreast cancer after low-dose irradiation. Lancet 1989;1:629–31.

47. Boyle P, Zheng T. Environmental factors influencing theepidemiology of breast cancer. Breast 2001;10(Suppl 3):1–8.

48. Laden F, Collman G, Iwamoto K, Alberg AJ, Berkowitz GS,Freudenheim JL, et al. 1,1-Dichloro-2,2-bis(p-chloropheny-l)ethylene and polychlorinated biphenyls and breast cancer:combined analysis of five US studies. J Natl Cancer Inst2001;93:768–76.

breast cancer control: signs of progress, but more work required

Documents